Search

The Kuhn-Munkres (KM) algorithm is a classical combinatorial optimization algorithm that is widely used for minimum cost bipartite matching in many real-world applications, such as transportation. For example, a ride-hailing service may use it to find the optimal assignment of drivers to passengers to minimize the overall wait time. Typically, given two bipartite sets, this process involves computing the edge costs between all bipartite pairs and finding an optimal matching. However, existing works overlook the impact of edge cost computation on the overall running time. In reality, edge computation often significantly outweighs the computation of the optimal assignment itself, as in the case of assigning drivers to passengers which involves computation of expensive graph shortest paths. Following on from this observation, we observe common real-world settings exhibit a useful property that allows us to incrementally compute edge costs only as required using an inexpensive lower-bound heuristic. This technique significantly reduces the overall cost of assignment compared to the original KM algorithm, as we demonstrate experimentally on multiple real-world data sets, workloads, and problems. Moreover, our algorithm is not limited to this domain and is potentially applicable in other settings where lower-bounding heuristics are available.

This paper presents small-signal modeling for state-of-the-art radio-frequency (RF) silicon-on-insulator (SOI) metal–oxide–semiconductor field effect transistors (MOSFETs). Especially, we have incorporated the neutral-body effect in our RF SOI model. This effect is significant in both RF extrinsic and intrinsic modeling stages. In addition, we have developed a physically-accurate parameter extraction method based on our analytical expressions. Our modeling results agree well with the measured data and can capture the frequency dependences of both output conductance and capacitance in the GHz frequency region. The anomalous S22 and S21 behaviors as well as the output conductance rising effect observed in our measurements can be predicted and described using the proposed model.

Hot-carrier (HC) effects on high-frequency and RF power characteristics of Si/SiGe HBTs are investigated in this paper. By using the two-tone load-pull measurement, we find that not only the cutoff frequency, but also the output power performances of Si/SiGe HBTs are suffered by the HC stress. In this work, S-parameters and intrinsic elements of an equivalent hybrid-/spl pi/ model were used to validate the HC effects on high-frequency characteristics. With different bias conditions, the degradations of cutoff frequency, power gain, and linearity are found to be worse under constant base-current measurement than that under constant collector-current measurement. The HC-induced degradations on the current gain, transconductance, and ideality-factor of base and collector currents are analyzed to explain the experimental observations.

Passive voltage contrast (PVC) has traditionally been used by semiconductor engineers for end-of-line post-mortem analysis. PVC distinguishes between open and short structures and is both nondestructive and noncontact. When applied during process development for in-line characterization, it allows wafers to be examined at multiple points, where electrical probing might not be feasible. This provides feedback on the cumulative effect of the process on critical parameters such as oxide integrity and can reduce development cycle times because wafers do not have to be deprocessed in order to determine the exact location of failures. Two case studies are presented in this article, demonstrating the use of PVC in a process development environment.

For the first time, the temperature dependences of radio frequency (RF) metal oxide semiconductor field effect transistors' intrinsic noise currents, including the induced gate noise current (ig), channel noise current (id) and their correlation noise current, are experimentally investigated. The power spectral densities for the induced gate noise current and correlation noise current are found to rise as temperature increases, and decline for the channel noise current. Moreover, by using van der Ziel's noise model, our experimental results show that, besides ambient temperature, the channel conductance is the main factor dominating the RF noise behaviors. Finally, bias dependence results are also presented.

We have investigated the radio frequency (RF) extrinsic resistance extraction for partially-depleted (PD) silicon-on-insulator (SOI) metal-oxide-semiconductor field effect transistors (MOSFETs). Although the thick buried oxide in SOI devices can block the substrate coupling, the SOI neutral-body coupling effect is significant for RF applications. An equivalent circuit considering this effect has been proposed. Based on this equivalent circuit, a new model capturing the frequency dependence of extrinsic resistances has been derived. After considering the impact of quasi-neutral body, we have developed a physically accurate RF extrinsic resistance extraction methodology for PD SOI MOSFETs

Two K-Band low-noise amplifiers (LNAs) are designed and implemented in a standard 0.18 /spl mu/m CMOS technology. The 24 GHz LNA has demonstrated a 12.86 dB gain and a 5.6 dB noise figure (NF) at 23.5 GHz. The 26 GHz LNA achieves an 8.9 dB gain at the peak gain frequency of 25.7 GHz and a 6.93 dB NF at 25 GHz. The input referred third-order intercept point (IIP3) is >+2 dBm for both LNAs with a current consumption of 30 mA from a 1.8 V power supply. To our knowledge, the LNAs show the highest operation frequencies ever reported for LNAs in a standard CMOS process.

Very low power loss /spl les/0.6 dB at 110 GHz and noise of

High quality RF inductors, very low loss and noise CPW and microstrip lines, advanced broad and narrow band filters, and ring resonators have been achieved on Si substrates, using an optimized proton implantation process. The RF performance up to 100 GHz is close to that for ideal devices designed by EM simulation for lossless substrates.

This paper proposes a cost-effective RF power cell manufactured in an advanced 0.13 um CMOS technology. Without adding additional masks, cost, and process, the power performance can be improved just by using the standard N-well and shallow-trench-isolation processes to form a higher resistive region. This ldquoPseudo-Drainrdquo structure increases the breakdown voltage to more than 4.3V and is higher than the value of 2.5V of the standard 0.13 um core-MOS transistor. This transistor exhibits a high fTtimesBVDS product of 352 for CMOS power FETs. Cutoff frequency and maximum oscillation frequency of 83 GHz and 124 GHz were achieved at a drain bias of 1.2V, respectively; while the maximum power gain, output power and power-added efficiency were 25.6 dB, 19 dBm, and 55%, respectively. Good RF linearity and noise figure were also obtained, as demonstrated by an OIP3 and NFmin of 28.32 dBm and 0.4 dB. The presented RF power transistor is cost effective and can be used for power amplifier integration in RF-CMOS SOC.

Emissions of reduced sulphur compounds (RSCs) from the primary and secondary clarifiers at a Kraft mill were measured for respectively 8 and 22 days using a floating flux chamber. In the primary clarifier, dimethyl disulphide (DMDS) had the highest mean flux (0.83 microg s(-1) m(-2)) among all RSCs, and the mean flux of total reduced sulphur (TRS) was 1.53 microg s(-1) m(-2). At the secondary clarifier, dimethyl sulphide (DMS) had the highest mean flux (0.024 microg s(-1) m(-2)), and the mean flux of total reduced sulphur (TRS) was 0.025 microg s(-1) m(-2). Large variations in fluxes as a function of sampling date were observed in both clarifiers. Emission fluxes of DMS from the secondary clarifier were correlated with temperature in the flux chamber and with the biological and chemical oxygen demands (BOD and COD) of the wastewater. Emission rates of RSCs from the clarifiers were found to be insignificant by comparison with other mill sources.

This paper addresses the variations that presently exist regarding the definition, determination, and reporting of detection limits for volatile sulphur compounds by gas chromatography with pulsed flame photometric detection (GC-PFPD). Gas standards containing hydrogen sulphide (H(2)S), carbonyl sulphide (COS), sulphur dioxide (SO(2)), methyl mercaptan (CH(3)SH), dimethyl sulphide (DMS), carbon disulphide (CS(2)), and dimethyl disulphide (DMDS) in concentrations varying from 0.36ppb (v/v) up to 1.5ppm (v/v) in nitrogen were prepared with permeation tubes and introduced in the gas chromatograph using a 0.25-ml gas sampling loop. After measuring the PFPD response versus concentration, the method detection limit (MDL), the Hubaux-Vos detection limit (x(D)), the absolute instrument sensitivity (AIS), and the sulphur detectivity (D(s)) were determined for each sulphur compound. The results show that the MDL determined by the US Environmental Protection Agency procedure consistently underestimates the minimum concentrations of volatile sulphur compounds that can be practically distinguished from the background noise with the PFPD. The Hubaux-Vos detection limits and the AIS values are several times higher than the MDL, and provide more conservative estimates of the lowest concentrations that can be reliably detected. Sulphur detectivities are well correlated with AIS values but only poorly correlated with MDL values. The AIS is recommended as a reliable and cost-effective measure of detection limit for volatile sulphur compounds by GC-PFPD, since the AIS is easier and faster to determine than the MDL and the Hubaux-Vos detection limit. In addition, this study confirmed that the PFPD response is nearly quadratic with respect to concentration for all volatile sulphur compounds.

The extraction of extrinsic resistances is essential to RF CMOS modeling. In this work, we investigate the extrinsic resistance extraction for PD SOI MOSFETs. We have shown that, for RF SOI MOSFETs, the coupling path between the source and drain terminals through the neutral-body region beneath the gate-oxide layer makes the resistance expressions behave frequency-dependently. After taking this effect into account, we develop a physical RF extrinsic resistance extraction methodology for PD SOI MOSFETs

This paper investigates hot-carrier (HC) effects on the RF power and linearity characteristics of MOS transistors using load-pull measurement. We found that the RF power characteristics are affected by the HC stress, and the linearity of MOS transistors is clearly degraded after HC stress at constant gate voltage measurement. However, at high gate voltage bias, the HC-induced power degradation is much reduced compared with that under low gate voltage regimes. In addition, HC effects on linearity can be softened by biasing the transistor at constant drain currents. These experimental observations can be explained by the change of threshold voltage, transconductance, subthreshold swing, and mobility degradation coefficient under HC stress.

Strong dependence of finger number on minimum noise figure (NF/sub min/) is observed in 0.18 /spl mu/m MOSFETs. A lowest NF/sub min/ of 0.93 dB is measured at 5.8 GHz using 50 fingers but increases as either increasing or decreasing finger number. We have used a self-consistent S-parameter and NF/sub min/ model to analysis this abnormal finger number dependence, and the reason is due to the combined effect of reducing gate resistance and increasing substrate loss as increasing finger number.

As scaling down the RF MOSFET from 0.18 to 0.13 /spl mu/m technology nodes, the f/sub T/ increases but the NF/sub min/ becomes worse by increasing /spl sim/0.2 dB. A small NF/sub min/ of 0.93 dB is measured at 5.8 GHz in 0.18 /spl mu/m MOSFET using 50 fingers but increases as either increasing or decreasing finger number. This abnormal dependence and higher noise at 0.13 /spl mu/m is accurately analyzed by equivalent circuit model and due to the combined gate resistance and substrate effect.

Phase-only filters that maximize the output signal-to-noise ratio (for colored noise input) are presented. These optimal POFs are then tested with the help of simulations to illustrate their ability to detect targets in natural backgrounds.